Cryptographer Matthew Green Warns AI Could Undermine Public-Key Encryption Standards
Cryptographer Matthew Green warned that rapid AI advancements could outpace the human ability to update cryptographic standards, estimating a 15% chance of losing functional confidence in existing public-key encryption algorithms. He emphasized that surviving AI-driven cryptographic breakthroughs requires advance preparation, as standards updates move orders of magnitude slower than AI capabilities. Modern digital security relies heavily on public-key cryptography for sensitive operations ranging from web browsing to financial transactions. If AI accelerates vulnerability discovery faster than standards bodies can test and deploy replacements, global security infrastructure could face unprecedented disruption. Green assigned a 1% probability to living in "Minicrypt"—a computational world where public-key encryption is fundamentally impossible—and a 15% chance of practical loss of confidence in current public-key algorithms. He argued that even leveraging AI to assist human standardization will not close the speed gap with AI-driven cryptanalysis.
## BACKGROUND
Formulated by computer scientist Russell Impagliazzo in 1995, the "Five Worlds" framework categorizes potential scenarios for computational complexity and cryptography. In "Minicrypt," one-way functions exist—enabling symmetric encryption and hashing—but secure key exchange over open channels via public-key cryptography is mathematically impossible. Public-key encryption relies on difficult mathematical problems, which could become vulnerable if automated tools discover unexpected algorithmic shortcuts.